ViSA: Visual and Semantic Alignment for Robust Scene Text Recognition.

Zhenru Pan Zhilong Ji Xiao Liu Jinfeng Bai Cheng-Lin Liu

Published in: ICDAR (2) (2023)

Keyphrases

scene text recognition
neural network
visual information
semantic content
data sets
high level
domain specific
visual features
semantic representation
semantically relevant
metadata
low level
computationally efficient
visual cues